Agent Features How to Use Pricing App & API FAQ Generate Free →
◆ Text to 3DMeshy 7.1

Meshy Text to 3D: Type an Idea, Get a Working 3D Model

No sketch, no reference photo just a sentence. Meshy's Text to 3D turns a written prompt into a textured, PBR-ready mesh in about a minute, on the current Meshy 7.1 engine. Here's how the prompt actually becomes the model, and where it still falls short.

Generate Your First Model Free → See the Honest Section

~60 second generations · up to 8K PBR textures · exports to 8 formats



Latest updates

What's new in Text to 3D

Sep 2026

Meshy 7.1 is the default engine

Text to 3D now runs on Meshy 7.1 by default sharper surface detail and better proportion accuracy than Meshy 6. Meshy 6 stays selectable for anyone mid-project.

Ongoing

Textures up to 8K

PBR texture maps now scale from 2K up to 8K resolution depending on plan and settings see how texturing works.

Ongoing

Smart Topology option

Choose Standard for max detail or Smart Topology for lean, game-ready meshes at generation time no separate remesh pass required.

Jun 2026

3D Agent for batches

Need a whole prop list from text, not one model at a time? The 3D Agent chains multiple Text to 3D calls in one chat see a full set in one chat.

Ongoing

Export list unchanged

FBX, OBJ, GLB, USDZ, STL, BLEND, 3MF and DXF remain available on every generation see features.

Ongoing

Credit cost stable

Each Text to 3D generation is still 20 credits, successful or not see pricing.



The process

How a sentence becomes a mesh

Write the prompt

Describe the object, style and material the more specific, the better the first result.

Set the options

Pick pose (None/A-Pose/T-Pose/Custom), model type (Standard/Smart Topology), and symmetry.

Generate (~60s)

Meshy returns an untextured mesh first, then applies full PBR texturing in the same pass.

Refine

Not quite right? Re-prompt, adjust the seed, or switch to Image to 3D with a reference.

Export or continue

Download in your target format, or send straight into free auto-rig and animation.



Prompt quality = output quality

Weak prompt vs strong prompt

! Weak: "a chair"

  • No material, no style, no era Meshy guesses everything
  • Generic proportions, ambiguous texture pass
  • Higher odds you'll spend another 20 credits re-rolling

✓ Strong: "a weathered oak dining chair, carved backrest, medieval tavern style, worn leather seat"

  • Material (oak, leather), style (medieval tavern) and detail (carved backrest) all specified
  • Texture pass has clear direction wood grain, leather wear
  • Far more likely to be usable on the first generation
[subject] + [material/texture] + [style/era] + [notable detail] → e.g. "a cracked stone gargoyle, gothic cathedral style, mossy weathering, wings folded"


Real prompts

What people actually type

"low-poly pine tree, stylized game asset" "sci-fi energy pistol, glowing blue accents" "rustic wooden barrel, iron bands, weathered" "cute cartoon fox, big eyes, soft fur texture" "ornate treasure chest, gold trim, fantasy RPG" "modern minimalist lamp, matte ceramic base" "medieval shield, battle-worn, family crest" "cyberpunk street vendor cart, neon signage"


Quick facts

What Text to 3D is built for

Current engineMeshy 7.1 (Meshy 6 still selectable)
Pose controlNone, A-Pose, T-Pose, Custom
Mesh detailUp to ~600K faces
TexturingFull PBR set, up to 8K
Export formatsFBX, OBJ, GLB, USDZ, STL, BLEND, 3MF, DXF
Generation time~60 seconds
🏆

Excellent for

Game props, original characters, sculptural and decorative objects, fantasy/sci-fi assets.

🎯

Good with care

Product concepting, architectural elements works, but benefits from tighter, more technical prompts.

📐

Better with a reference

Matching an exact real object? Image to 3D with up to 4 angle photos beats text alone.



The honest section

What Text to 3D still gets wrong

! Known limitations



The numbers

What generations cost in practice

PlanCredits/mo≈ GenerationsCost
Free100~5$0
Pro1,000~50$20/mo ($10 first month)
Premium3,000~150See pricing

Every generation is 20 credits whether or not you keep it see full pricing and discounts for cheaper ways in.



Real output

Prompts, turned into models



See it work

Prompt to model, start to finish

Text to 3D, texturing and export

Prompt-writing walkthrough



Text to 3D FAQ

Common questions

How does Meshy convert text to 3D?

Meshy's model interprets your prompt, generates an untextured mesh, then applies full PBR texturing in the same pass typically around 60 seconds on the current Meshy 7.1 engine.

Is there a free trial?

Yes. The free plan gives 100 credits/month (about 5 generations) with no card required.

Can I type a copyrighted character?

Generation is technically permitted, but commercial ownership and distribution rights are not granted for outputs based on someone else's IP see license and commercial use.

How much does one generation cost?

20 credits per generation, whether or not you keep the result. Budget for retries when writing vague prompts see pricing.

What's the difference between Text to 3D and Image to 3D?

Text to 3D starts from a written description; Image to 3D starts from up to 4 reference photos and generally produces more accurate proportions for a specific real object.

Can I animate what I generate?

Yes. Auto-rigging and 500+ animation presets are free (0 credits) on every plan see features.

What export formats are supported?

FBX, OBJ, GLB, USDZ, STL, BLEND, 3MF and DXF covering game engines, DCC tools and 3D printing slicers.



Ready to type your first prompt?

100 free credits, no card required see what your idea looks like as a textured 3D model in about a minute.

Generate Your First Model Free →